You are on page 1of 14

14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Received: 17 December 2020 Accepted: 18 November 2022 Published online: 19 January 2023

DOI: 10.1002/agj2.21264

ORIGINAL ARTICLE
A g ro n o m y, S o i l s , a n d E nv i ro n m e n t a l Q u a l i t y

Impacts of long-term fertilization on crop yield and microbial


communities under rice–rapeseed rotation

Jifu Li Guoyu Gan Qianqian Zeng

Engineering Research Center of Ecology


and Agricultural Use of Wetland, Ministry Abstract
of Education/College of Agriculture, Soil microbes play an integral role in agricultural production and soil fertility, while
Yangtze University, Jingzhou, China
limited information is available about the effects of long-term fertilization on rice
Correspondence (Oryza sativa L.)–rapeseed (Brassica napus L.) rotation yield, soil properties, and
Li Jifu, No. 266, Jingmi Road, Jingzhou microbial communities in central China. Here, we report the response of plant–soil
District, Jingzhou City 434023, Hubei
Province, China. system to different mineral fertilizer inputs, including the balanced fertilization with
Email: jifuli@yangtzeu.edu.cn N, P, and K (NPK) fertilizers and unbalanced fertilizers without one of the major
nutrients (PK, NK, and NP) for 10 years. Compared with the NPK treatment, the
Assigned to Associate Editor Xinda Lu.
average annual decrease rates of rice and rapeseed yield in the PK, NK, and NP
Funding information treatments were 12.5%, 18.1%, 6.4% and 33.1%, 54.8%, 16.0%, respectively. Long-
Engineering Research Center of Ecology
term fertilization significantly changed the soil chemical properties, especially the
and Agricultural Use of Wetland, Ministry
of Education, Grant/Award Number: pH and available K content. Compared with the NPK fertilization, the NK treatment
KF2020015; Key Laboratory of Fertilizer increased the soil microbial composition and diversity, whereas PK and NP treat-
Utilization, Ministry of Agriculture and
Rural Affairs, China, Grant/Award Number:
ments had a reverse impact. Available K, pH, and slowly available K were closely
KLFAW201901 related to alterations of the dominant bacterial community. On the contrary, the
Olsen-P, slowly available K, and pH were significantly correlated with the fungal
community. Soil properties changed by fertilization had a more direct effect on crop
yield than the microbial community; moreover, the bacterial group had a higher influ-
ence than the fungal group on crop yield. Therefore, a balanced nutrient input using
NPK fertilization is an effective management strategy for soil biological fertility and
sustainable crop yield.

1 INTRODUCTION example, soil acidification, nutrient loss, and land capacity


degeneration (Lu et al., 2018; Reed, 2017; Zhang et al., 2020).
Mineral fertilizer, especially nitrogen (N), phosphorus (P), As a vital component of the soil ecosystem, microorganisms
and potassium (K), is a large contributor to global food pro- play a critical role in the formation of soil fertility and the bio-
duction. Reasonable fertilization can promote the positive geochemical cycle of nutrients (Ullah et al., 2020). The effect
development of the soil ecosystem; however, inappropriate of fertilization on soil microorganisms is different due to the
or excessive fertilization will cause production reduction and quantity, type, time of fertilization, and the combination of
a series of environmental problems in agro-ecosystem, for different types of fertilizers. Meanwhile, soil microorganisms
respond to different fertilizer rates, leading to changes in the
Abbreviations: ACE, abundance-based coverage estimator; GOF, biomass, activity, and community structure (Ling et al., 2017;
goodness of fit; OTU, operational taxonomic units; SOM, soil organic Wang et al., 2018; Wasaki et al., 2018).
matter.

© 2022 The Authors. Agronomy Journal © 2022 American Society of Agronomy.

208 wileyonlinelibrary.com/journal/agj2 Agronomy Journal. 2023;115:208–221.


14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 209

Previous studies have shown that fertilization could alter


the nutrient content of soil (i.e., total N, available P, and Core Ideas
available K) and directly drive the evolution of soil microbial
communities (Cookson et al., 2005; Mooshammer et al., ∙ Crop yields were sensitive to P deficiency than N
2014). However, other studies believe that fertilization indi- and K deficiencies.
rectly changes soil properties. Acting on soil microorganisms, ∙ Coupling interaction existed between soil proper-
Zeng et al. (2016) found that the decrease in soil pH caused ties and microbes on yield.
by the application of N fertilizer was the primary cause of ∙ The responses of microbial diversity to P, K, and N
bacterial community changes. Long-term P fertilizer input deficiency were different.
increased soil microbial P immobilization by decreasing the ∙ Available K, pH, and slowly available K were
relative abundance of the P-starvation response gene (phoR) closely related to bacterial community.
and increasing the relative abundance of the low-affinity ∙ Olsen-P, slowly available K, and pH were signifi-
inorganic phosphate transporter gene (Dai et al., 2020). In cantly correlated with fungal community.
black soil, Proteobacteria, Acidobacteria, and Actinobacteria
were the dominant bacteria when different amounts of K
fertilizer were applied, while the relative abundance of Aci-
dobacteria was significantly decreased with the increase of 2 MATERIALS AND METHODS
K application (Yu et al., 2019). Fan et al. (2018) showed that
some keystone species of bacteria had higher functional gene 2.1 Experimental site description
copies associated with N and P element cycling in controlling
soil function and crop production. Meanwhile, the microbial A long-term field trial was initiated in 2011 at the Shuangdian
community was highly responsive to K fertilization, which Experimental Station located in Jingzhou City, Jianghan Plain
was associated with lower upland crop yield and higher (N 30˚33′24″, E 112˚4′56″, altitude 52 m), Hubei Province,
abundance of potential fungal pathogens. However, the indi- central China, adopting a middle rice–rapeseed (Brassica
vidual effects of N, P, and K (NPK) fertilizers and balanced napus L.) rotation system developed in 1999 (Tang et al.,
fertilizer inputs on bacterial and fungal communities, and 2018). Monthly temperature and rainfall during the exper-
the potential effects of fertilization-induced microbial groups iment (2011–2020) are shown in Figure S1. The average
on rotation system productivity in paddy soils have not been annual rainfall was 1140 mm, and the air temperature was
addressed. 15˚C. Soil type was classified as Alfisol using the Word Soil
Rice (Oryza sativa L.)-based cropping system is the Classification of the Food and Agriculture Organization (sand
most widely distributed and productive agroecosystems in 26.3%, silt 40.8%, and clay 32.5%). At the beginning of the
Southern China, Southeast Asia, and South Asia. The joint experiment (June 2011), the selected soil basial properties at a
contribution rate of NPK fertilizers application to crop yield depth of 0–20 cm were as follows: pH 6.4, soil organic matter
was more than 50% in China (Tian et al., 2018). Eo and Park (SOM) 28.8 g kg−1 , total N 0.6 g kg−1 , available N 275.3 mg
(2016) illustrated that the application of N fertilizer or P fer- kg−1 , Olsen-P 4.44 mg kg−1 , available K 96.6 mg kg−1 , and
tilizer alone reduced the bacterial diversity and simplified slowly available K 528.2 mg kg−1 . According to the nutrient
the interactions among dominant bacterial members. With deficiency index of paddy fields, the available nitrogen con-
the project on soil testing for formulated fertilization in a tent for the field experiment is high, indicating that the field
pilot area carried out in 2005, balanced fertilization has been soil is rich in N and K but deficient in P (Tang et al., 2018).
widely accepted by farmers, but insufficient or excessive fer-
tilization also took place in rice region of the Yangtze River
Basin. There is not enough information on the effects of indi-
2.2 Experiment design
vidual nutrient input deficiency stress (N, P or K fertilizer) and
balanced fertilization on soil biological fertility; especially,
A complete randomized block design was conducted with four
there is a lack of long-term field comparative studies on paddy
treatments and three replications. The treatments were as fol-
soil (Radhakrishnan & Varadharajan, 2016). Therefore, we
lows: (1) NPK: balanced fertilizers of NPK were applied; (2)
carried out this experiment to assess the relationships between
PK: mineral fertilizers of P and K were applied; (3) NK: min-
soil properties, microbial community altered by fertilization
eral fertilizers of N and K were applied; and (4) NP: mineral
with crop yield.
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
210 LI ET AL.

fertilizer of N and P were applied. The regional balanced fer- Available K and slowly available K were extracted with 1 mol
tilizer application for annual rice and rapeseed season were L−1 NH4 OAc and 1 mol L−1 HNO3 solution, respectively, and
180 kg N ha−1 , 90 kg P2 O5 ha−1 , 120 kg K2 O ha−1 , and measured by a photoelectric flame photometer. The SOM was
1.6 kg B ha−1 only for rapeseed season (Yousaf et al., 2017). determined using a wet oxidation procedure with potassium
During rice season, N (urea, 46% N) was applied in three dichromate (K2 Cr2 O7 )-sulfuric acid (H2 SO4 ). Furthermore,
splits: 60% as basal fertilizer before rice transplanting, 20% soil available N was determined by the alkaline hydrolysis
at tillering stage, and 20% at the booting stage. K (potassium diffusion method.
chloride, 60% K2 O) was applied as 60% basal fertilizer and
40% booting fertilizer. P (superphosphate, 12% P2 O5 ) was
applied manually as basal fertilizers. During oilseed rape sea- 2.3.4 Soil DNA extraction and
son, N was applied in three splits: 60% as basal fertilizer, 20% high-throughput sequencing analysis
during the overwintering stage, and 20% at the beginning of
stem elongation. P, K, and B (sodium borate, 11% B) fertiliz- DNA was extracted from the soil samples (0.5 g) using a
ers were applied manually as basal fertilizers. After each crop Fast DNA Spin Kit for Soil (MP Biomedicals, USA) in accor-
was harvested, the crop straw was removed from the plots. dance with the protocol of the manufacturer. The quantity and
The specific fertilization time for a crop is shown in Figure quality of the DNA extracts were determined using a Nan-
S1. Local crop varieties, Eke No. 1 and Huashuang No. 11, oDrop 2000 spectrophotometer (Thermo Fisher Scientific).
widely planted in the experimental region with high yields The extracted DNA was stored at −20˚C for further analyses.
and extensive adaptability, were used for rice and rapeseed, An aliquot of the extracted DNA from each sample
respectively. The dimensions of the plot were 20 m2 , with a was used as the template for amplification. The V3–V4
length of 5.0 m and a width of 4.0 m. hypervariable regions of the bacterial 16S rRNA gene
sequences and the ITS region of the fungal rRNA gene
sequences were amplified (Wang et al., 2017). Amplicon
2.3 Methods libraries were prepared using tagged bacterial and fun-
gal universal primers, that is, 338F and 806R for bacte-
2.3.1 Crop harvest ria and ITS1F and ITS2R for fungi. The DNA samples
were amplified individually using the fusion primer pairs
For each season, the whole plots of mature rapeseed and rice 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R
plants were harvested and thrashed in each plot. The grains (5′-GGACTACHVGGGTWTCTAAT-3′) for bacteria and
were dried to determine the grain yield. ITS1F (5′-CTTGGTCATTTAGGAAGTAA-3′) and ITS2R
(5′-GCTGCGTTCTTCATCGATGC-3′) for fungi to generate
polymerase chain reaction (PCR) fragments. The following
2.3.2 Soil sampling and preparation
thermal program was used for amplification: initial denatura-
tion 98˚C for 2 min, followed by 27 cycles of denaturation at
On September 10, 2020 (1 day before rice harvest), soil sam-
98˚C for 15 s, annealing at 55˚C for 30 s, extension at 72˚C
ples were randomly collected from four points in each plot
for 30 s, and a final extension at 72˚C for 5 min. The PCR
between the rice lines at 0–20 cm depths by using an auger
reactions were performed in a 25 μl mixture containing 5 μl
with a diameter of 5.0 cm. Soil from the four core samples of a
of 5× reaction buffer, 5 μl of 5×GC buffer, 2 μl of dNTP
plot was mixed to obtain one composite sample. After remov-
(2.5 mM), 1 μl of forward primer (10 μM), 1 μl of reverse
ing stones, roots, and plant residue using a 2-mm mesh, each
primer (10 μM), 2 μl of DNA template, 8.75 μl of ddH2 O, and
sample was divided in half: one half was air-dried for physico-
0.25 μl of Q5 DNA polymerase (Tian et al., 2017). The PCR
chemical analyses, and the other half was immediately stored
products were purified using the AxyPrep DNA Gel Extrac-
at −80˚C for DNA extraction and microbiological analysis.
tion Kit (Axygen Biosciences) and quantified using Quantus
Fluorometer (Promega). The target sequences were performed
2.3.3 Determination of soil physicochemical on an Illumina MiSeq 250 sequencing platform by Shanghai
indexes Personal Biotechnology.

The air-dried soil samples were used to determine physico-


chemical properties according to the methods described by 2.3.5 Sequence processing
Ding et al. (2016). Briefly, soil pH was measured in water
(1:2.5 w/v) by a pH meter (PHS-3C, INESA Scientific Instru- In sequencing the original data to remove the primer adapter
ment, Shanghai, China). Olsen-P was extracted with 50 ml sequence, the processed low-quality bases were removed,
of 0.5 mol L−1 NaHCO3 (pH 8.5) and determined using an and then the remaining data were spliced to obtain valid
injection pump analyzer (AA3, Bran+ Luebbe, Germany). sequence data for each sample. Finally, using 97% as the
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 211

threshold, the 16S and ITS sequences were divided into oper- 54.8%, 16.0%, respectively. In the first three crop rotations,
ational taxonomic units (OTUs). Using QIIME 2 software, the the rice yield from the PK, NK, and NP treatments did not
UCLUST sequence comparison tool was used to cluster with show a significant reduction, whereas after the four rotation
97% sequence similarity and each sequence with the highest cycles, a significant decreasing tendency could be seen in
OTU degree was selected as the representative sequence of Figure 1a and Figure S2. The rice yield followed the order of
the OTU (Caporaso et al., 2010). For bacterial 16S rRNA and NPK > NP > PK > NK (Figure 1b). For the subsequent crop
fungal ITS genes, both the Greengenes database and the Silva rapeseed, a yield decrease effect appeared in the second rota-
database were used as template sequences for OTU classifica- tion, especially in the NK treatment and PK treatment, which
tion status identification (Edgar, 2010). After quality filtering are shown in Figure 1c and Figure S2.
and removal of chimeric sequences, 364,587 and 336,310
high-quality sequences were, respectively, clustered into
48,847 and 76,607 OTUs for each bacterial and fungal sample. 3.2 Soil properties

The results in Table 1 showed that among the four treatments,


2.4 Statistical analysis there were significant changes in the soil chemical properties.
Compared with the initial value, the pH for the NK, NP, and
The analysis of variance procedure in SPSS 16.0 (SPSS NPK treatments had a downward trend. The PK fertilization
Inc.) was used to perform data analysis on soil biogeo- had no significant difference from the initial value. Compared
chemical properties and alpha diversity. Differences at the with the initial value, the SOM content of each treatment
p < 0.05 level was considered statistically significant, as deter- increased to varying degrees, and the rate of increase in the
mined using the least significant difference test. The Chao1, NPK fertilization was the highest, reaching 16.3%. The results
abundance-based coverage estimator (ACE), and Shannon, of available N content showed that the content of the PK,
Simpson indexes were used to estimate the alpha diversity of NK, and NP treatments showed a downward trend, decreas-
each sample using MOTHUR (Schloss et al., 2009). To deter- ing by 12.4%, 4.7%, and 5.7%, respectively, while that of
mine the structural differences between bacterial and fungal the NPK fertilization increased by 4.0%. Compared with the
communities at different treatments, an analysis of similarities initial value, the NK treatment significantly reduced the soil
was also conducted using QIIME2 based on Bray–Curtis dis- Olsen-P content by 18.0%, and there was no significant differ-
tance measurements and abundance data (Wang et al., 2017). ence in other treatments. Ten years of NP fertilization led to
To determine which taxa were significantly affected, the linear a decrease in the available K content and the slowly available
discriminant analysis effect size algorithm was implemented K content in soil, which were 25.2% and 4.9% lower than the
(Segata et al., 2011). The “vegan” package in R 4.1.0 language initial values, respectively. In comparison, the NPK treatment
was used to perform similarity analysis, and redundancy increased the available K and the slowly available K contents
analysis (RDA) was used to access the effects of soil environ- by 14.7% and 1.4%, respectively.
mental factors on bacterial and fungal communities (Oksanen
et al., 2010). To reveal the potential pathways (soil proper-
ties, alpha diversity, and microbial community) that influence 3.3 Alpha diversity of the soil microbial
rice yield and rapeseed yield, partial least squares path models community
were evaluated using the goodness of fit (GOF) statistic (Bar-
beran et al., 2014) and assembled by the “inner plot” function The rarefaction curves (Figure S3) of all samples tended to be
using the “plspm” package of R 4.1.0. flat, suggesting that each sample was equally and reasonably
sampled. The coverage was over 0.99 in all samples, reveal-
ing that the sequencing depth was enough to capture almost
3 RESULTS all the diverse bacterial and fungal communities in this study
(Table 2). The richness and diversity of bacterial and fungal
3.1 Crop yield communities were different under various fertilization treat-
ments. The richness of bacteria and fungi (Chao1 and ACE)
Over the 10 years, crop yields were significantly affected were both NK > NPK > NP > PK fertilization, and there were
by different fertilization for rice and rapeseed (Figure 1). significant differences between treatments. The diversity of
The grain yields with NPK fertilizer were 8591 kg ha−1 soil bacteria and fungi (Simpson and Shannon) was ranked as
and 2325 kg ha−1 annual average for rice and rapeseed, NK, NPK > NP > PK fertilization and the differences were
respectively. Compared with the NPK treatment, the aver- significant. Therefore, compared with the NPK treatment, NK
age annual decrease rates of rice and rapeseed in the PK, treatment increased the soil microbial diversity, while the PK
NK, and NP treatments were 12.5%, 18.1%, 6.4% and 33.1%, and NP treatments did the opposite impact.
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
212 LI ET AL.

DIAGRAMA DE CAJAS Y BIGOTES

Valor máximo

Valor mínimo

Mediana

Media

Cuartil Q3

Cuartil Q1

Rango intercuartil

Valor atípico (outliers)

F I G U R E 1 Variation and distribution of grain yields for (a,b) rice and (c,d) rapeseed under different fertilization treatments PK, NK, NP, and
NPK from 2011 to 2020. The box plots in (b) and (d) display the distribution of rice and rapeseed yields, respectively, over the 10 years at different
treatments. A dot mark in the box indicates the mean value of crop yield. The upper, middle, and lower limits of each box denote, respectively, the
75th, 50th, and 25th percentages for crop yield, while different lowercase letters indicate significant differences in mean grain yield between
treatments at p < 0.05 in (b, d). *, **, *** indicate significant differences at p < 0.05, p < 0.01, p < 0.001 levels, respectively

TA B L E 1 Effects of long-term fertilization on soil chemical properties in bulk soils

Available N Olsen-P Available K Slowly available K


Treatment pH SOM (g−1 kg−1 ) (mg−1 kg−1 ) (mg−1 kg−1 ) (mg−1 kg−1 ) (mg−1 kg−1 )
PK 6.49 ± 0.28 a 29.5 ± 1.1 c 241.3 ± 28.5 c 4.23 ± 0.16 a 92.9 ± 6.3 b 520.6 ± 14.2 ab
NK 5.86 ± 0.17 b 31.1 ± 1.1 ab 263.2 ± 18.3 b 3.75 ± 0.90 b 98.0 ± 9.0 c 524.4 ± 15.4 a
NP 5.72 ± 0.13 bc 30.4 ± 0.9 b 259.7 ± 13.0 b 4.52 ± 0.23 a 72.3 ± 7.1 d 502.3 ± 9.7 b
NPK 5.68 ± 0.16 c 33.5 ± 1.2 a 286.2 ± 14.1 a 4.36 ± 0.38 a 115.8 ± 2.9 a 535.6 ± 34.2 a
Initial value 6.40 ± 0.19 a 28.8 ± 0.8 c 275.3 ± 10.7 ab 4.44 ± 0.49 a 96.6 ± 7.2 b 528.2 ± 21.5 a

Note: Means of three replicates are presented with standard deviation. Different letters in the same column indicate significant differences among four treatments and
initial values at the p < 0.05 level.
Abbreviation: SOM, soil organic material.

3.4 Composition of the soil microbial (Figure 2a), Proteobacteria and Acidobacteria had the highest
community relative abundance in each treatment, belonging to the pre-
dominant bacterial community (relative abundance > 10.0%),
Long-term fertilization altered the relative abundance of bac- with an average of 31.9% and 14.2%, respectively. The rela-
terial and fungal phyla in soil. Of the top 20 bacterial phyla tive abundances of Actinobacteria, Chloroflexi, Nitrospirae,
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 213

TA B L E 2 Richness and diversity of bacterial and fungal gene sequences in the soil samples

Richness index Diversity index


Treatment Chao1 ACE Simpson Shannon Coverage
Bacteria
PK 2225 ± 134 d 2138 ± 128 d 0.983 ± 0.059 a 8.62 ± 0.05 c 0.996 ± 0.052 a
NK 4215 ± 158 a 4187 ± 157 a 0.998 ± 0.030 a 10.78 ± 0.03 a 0.996 ± 0.031 a
NP 2894 ± 150 c 2822 ± 142 c 0.992 ± 0.052 a 9.31 ± 0.04 b 0.995 ± 0.052 a
NPK 3622 ± 172 b 3544 ± 163 b 0.997 ± 0.047 a 10.22 ± 0.04 a 0.994 ± 0.044 a
Fungi
PK 147 ± 9 d 138 ± 8 d 0.898 ± 0.038 b 4.95 ± 0.03 c 0.999 ± 0.042 a
NK 662 ± 25 a 645 ± 25 a 0.974 ± 0.037 a 6.68 ± 0.03 a 0.999 ± 0.038 a
NP 239 ± 11 c 223 ± 10 c 0.964 ± 0.036 a 5.91 ± 0.04 b 0.999 ± 0.045 a
NPK 553 ± 20 b 536 ± 28 b 0.978 ± 0.052 a 6.95 ± 0.04 a 0.999 ± 0.053 a

Note: Means of three replicates are presented with standard deviation. Different letters within one column indicate significant differences among four treatments at the
p < 0.05 level.
Abbreviation: ACE, abundance-based coverage estimator.

F I G U R E 2 The relative abundance of top 20 taxonomic groups for (a) bacteria and (b) fungi at the phylum level between different treatments
in 2020 and significant changes in (c) bacterial and (d) fungal key phylotypes under nutrient deficiencies identified using linear discrimination
analysis (LDA) linear discriminant analysis effect size
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
214 LI ET AL.

Latescibacteria, and Rokubacteria were higher, with aver- them, Anaeromyxobacter had the highest relative abundance
age of 6.1%, 2.3%, 1.7%, 1.4%, and 1.1%, respectively. In in the PK and NP treatments, and Haliangium and Candida-
the fungi phylum, Ascomycota was the dominant species, tus Solibacter (p < 0.05) had the highest relative abundance
with an average relative abundance of 49.7%, followed by in the NK fertilizer application.
Basidiomycota and Mortierellomycota with an average rela- At the genus level of fungi (Figure 3b), the dominant
tive abundance of 10.4% and 4.4%, respectively. Compared species were Fusarium and Mortierella, having an average rel-
with the NPK fertilization, Proteobacteria, Actinobacte- ative abundance of 5.4% and 4.4%, respectively, followed by
ria, Chloroflexi, Verrucomicrobia, and Planctomycetes in Xylodon, Lecanicillium, Umbilicaria, and Chaetomium, with
the PK, NK, and NP treatments all increased significantly, an average of 2.7%, 2.4%, 2.3%, and 2.0%. The relative abun-
with the increase rates being 8.1%–14.6%, 23.5%–50.9%, dance of Fusarium (p < 0.05) decreased significantly in the
18.4%–50.8%, 93.9%–414.0%, and 122.6%–396.4%, respec- PK and NP treatments than in the NPK treatment. Among all
tively (p < 0.05); however, Acidobacteria declined to varying treatments, Mortierella and Xylodon in the NK treatment had
degrees by 29.3%–51.4%. The relative abundances of Bac- the highest relative abundance, while Lecanicillium, Umbili-
teroidetes, Gemmatimonadetes, and Firmicutes were different caria, and Chaetomium (p < 0.05) had the lowest relative
in the NK, PK, and NP treatments. Compared with the NPK abundance.
fertilization, Bacteroidetes, Gemmatimonadetes, and Firmi-
cutes in the NK fertilization significantly increased (p < 0.05)
by 55.7%, 14.4%, and 43.3%, respectively. At the same time, 3.5 Beta diversity of the soil microbial
Bacteroidetes, Gemmatimonadetes, and Firmicutes in the PK community
and NP treatments decreased at an average of 57.9%, 88.8%,
and 34.6%, respectively. The Venn diagram (Figures 4) shows that the microbial pop-
As for the fungal community (Figure 2b), compared with ulation had both shared components and unique parts. The
the NPK treatment, the relative abundance of Ascomycota PK, NK, and NP treatments shared 3.92%, 13.76%, and 7.04%
significantly reduced by 17.5% (p = 0.023), and the relative of the bacterial OTUs with the NPK treatment, while the
abundance of Basidiomycota and Mortierellomycota signif- unique OTUs of the PK, NK, NP, and NPK treatments were
icantly increased by 175.3% and 69.0%, respectively, in the 12.36%, 28.16%, 15.81%, and 20.69%, respectively. The PK,
NK treatment. The relative abundances of Basidiomycota, NK, and NP treatments shared 3.38%, 13.19%, and 5.73%
Mortierellomycota, and Rozellomycota in the NP treatment of fungal OTUs with the NPK treatment, while the unique
were reduced by 17.7%, 45.6%, and 77.2%, respectively OTUs of the PK, NK, NP, and NPK treatments were 6.67%,
(p < 0.05), compared with those of the NPK treatment. In 36.5%, 10.52%, and 27.63%, respectively. Additionally, prin-
comparison, Olpadiomycota and Glomeromycota in the PK cipal coordinate analysis showed that bacterial communities
treatment increased by 287.3% and 2428.6%, respectively. in the NK and NPK treatments were distinct from those in
At the level of bacterial phylum, Proteobacteria, Rokubac- the PK and NP treatments along the x-axis (Figure 4c), and
teria, and Planctomycetes were found to be the significantly the first principal component (x-axis) accounted for 72.2% of
altered taxa in the PK fertilization (Figure 2c). Actinobacteria, the total variation. Similarly, the first two principal coordi-
Gemmatimonadetes, Bacteroidetes, Chloroflexi, Verrucomi- nates represented 79.0% of the variation in fungi (Figure 4d)
crobia, and Firmicutes were found to be sensitive to the communities, in which the first principal component (x-axis)
NK fertilization. Interestingly, the predominant Nitrospi- accounted for 50.6% of the total variation.
rae, Latescibacteria, and Patescibacteria were significantly To further compare the differences in the microbial com-
altered in the NP fertilization, and only Acidobacteria positions of the four treatments, a heatmap was used to
was positively altered in the recommended NPK fertilizer demonstrate the genus composition. The distribution of domi-
application. At the level of fungal phylum (Figure 2d), nant bacteria in each treatment was well marked; specifically,
Glomeromycota, Olpidiomycota, and Rozellomycota were the in the NK treatment, up to 16 species could be identified in
significantly altered taxa in the PK treatment. Basidiomycota, the number of abundant bacteria (Figure 5). On the contrary,
Mortierellomycota, Zoopagomycota, and Chytridiomycota the abundant bacteria in the PK, NP, and NPK treatments
were sensitive to the NK treatment. Conversely, Ascomycota, were 10, nine, and five types of species, respectively. Further-
Entorrhizomycota, and Kickxellomycota clearly changed in more, the cluster analysis results reflected that the PK and NP
the NP treatment. Mucoromycota and Aphelidiomycota were treatments were similar, and so were the NK and NPK treat-
altered in noticeable manner in the NPK treatment. ments (Figure 5a). The results of the fungal genus also showed
At the bacterial genus level (Figure 3a), Anaeromyxobacter that the distribution and relative abundance of the dominant
was the dominant genus, with an average relative abundance fungal groups in each treatment were significantly different
of 6.27%, followed by Geobacter, Haliangium, Candidatus (Figure 5b). At the same time, the cluster analysis also indi-
Solibacter, and Nitrospira, with an average relative abundance cated that the NK and NPK treatments were similar and the
of 0.74%, 0.73%, 0.65%, and 0.51%, respectively. Among PK and NP treatments were homogeneous.
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 215

F I G U R E 3 The relative abundance of preponderant taxonomic groups for (a) bacteria and (b) fungi at the genus level between different
treatments in 2020. The data represent the mean values of the three replications

F I G U R E 4 Venn graph of (a) bacterial and (b) fungal operational taxonomic units from soil samples and the principal coordinate analysis
(PCoA) plot of the dissimilarity between the treatments for (c) bacteria and (d) fungi based on Bray–Curtis differences in 2020

3.6 Correlation of dominant microbial (Figure 6), indicating that long-term unbalanced fertilization
communities with soil properties had caused the differentiation of the soil microbial commu-
nity (bacteria p = 0.002 and fungi p = 0.002). The available
The dominant microbial groups of four treatments were K, pH, slowly available K, SOM, and Olsen-P had a notewor-
separated from each other in the two-dimensional sorting thy impact on the bacterial community, which explained the
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
216 LI ET AL.

F I G U R E 5 The heatmap displays the dominant (a) bacteria and (b) fungi variances among the treatments in 2020 at the genus level. The red
color represents genera with higher relative abundance in the corresponding sample, and the green color represents genera with lower relative
abundance

F I G U R E 6 Redundancy analysis (RDA) demonstrate the relationships between soil property indexes (red arrow) and the dominant (a)
bacterial and (b) fungal community (blue arrow) among soil samples. The direction of arrows indicates correlation with the first two canonical axes
and the length of arrows represents the strength of the correlations. SOM, soil organic matter

variation by 53.5%, 23.1%, 11.6%, 6.9%, and 4.8%, respec- 3.7 Potential pathways influencing crop
tively. Also, the Olsen-P, slowly available K, pH, available K, yield
and SOM had a prominent impact on the fungal community,
and the variations were explained by 61.2%, 17.7%, 9.2%, The results showed that the final model had a GOF of 0.73
8.6%, and 3.2%, respectively. and 0.56 for the bacterial and fungal groups, respectively
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 217

F I G U R E 7 Partial least squares path models for the rice and rapeseed yield in 2020. (a,b) Arrows indicate causal relationships, supplemented
by a path coefficient, and continuous and dash lines indicate positive and negative relationships. (c,d) The amount of the variability is explained by
the variables. Path coefficients are calculated after 1000 bootstraps. BC1, Bray–Curtis PCoA1; BC2, Bray–Curtis PCoA2; GOF, goodness of
fit; SOM, soil organic matter; UP1, unweighted UniFrac PCoA1; UP2, unweighted UniFrac PCoA2; WP1, weighted UniFrac PCoA1; WP2,
weighted UniFrac PCoA2

(Figure 7). The pathways of soil properties, alpha diver- is generally manifested as N > P > K and rapeseed is more
sity, and bacterial community composition together explained sensitive to nutrient deficiency stress than rice in the same
72.8% of the total variance in rice and rapeseed yield, whereas site, which is consistent with the contribution rate of fertilizer
those three items represented 41.3% of the variation in the (Chen et al., 2020). However, our results showed P > N > K
fungal group. This suggested that soil proprietary (path coeffi- fertilizer (Figure 1), which may be related to the lower avail-
cient = 0.79) had a greater impact on crop yield than microbial able P content in this region compared to available N and
properties, especially the bacterial group had more influence available K content (Table 1). However, compared to rice
than the fungal group on crop yield. season, soil moisture and temperature conditions reduce the
lower P bioactivity in the rapeseed season (Li & Zhong, 2021).
From the dynamic changes of crop yield in 10 years, the
4 DISCUSSION decrease rate of the NP treatment was slightly better than that
of the NK and PK treatments in the paddy-upland rotation
4.1 Effects of long-term fertilization on compared to that of the NPK treatment, and the drop ten-
rotation crop yields dency was noticeable in the rapeseed season. Compared with
the initial soil properties, the available K content in the end of
Reasonable application of mineral fertilizers is necessary experiment was higher than the content of available N and
for enhancing crop production and sustainable soil fertility. available P. With the extension of planting years, the sup-
Related studies have shown that the individual contribution ply capacity of soil fertilizer gradually decreased, which will
rate of NPK fertilizers on the yield of rice and rapeseed lead to a reduction in crop yield (Tang et al., 2018; Yousaf
was 20.3%, 12.2%, and 8.2% and 41.9%, 21.4%, and 11.5%, et al., 2017). Moreover, the results shown in Figure 7 indi-
respectively (Bai et al., 2015; Li et al., 2017; Yan et al., 2019). cate that the soil properties had more direct effect than the
Therefore, the nutrient limiting factor sequence of rice field community composition and the alpha diversity of bacterial
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
218 LI ET AL.

and fungal groups on crop yield. Therefore, nutrient uptake tinuous application of P fertilizer in eucalypts forest. There
by rice mainly comes from the soil stores, whereas rapeseed were also studies suggesting that long-term application of P
depends more on exogenous NPK fertilizers (Li et al., 2017; fertilizer increased the diversity of soil bacteria (Ling et al.,
Zhang et al., 2020). 2017; Wasaki et al., 2018).
The diversity of bacteria and fungi in the PK treatment
was the lowest (Table 2), indicating that the soil available N
4.2 Effects of long-term fertilization on soil content significantly affects the vitality of microorganisms.
chemical properties Therefore, for soil microorganisms, soil N availability is still
the primary limiting factor (Miransari, 2011; Zeng et al.,
The results demonstrated that after 10 years of fertilization, 2016). The correlation analysis of soil properties and diver-
the initial values of the physiochemical properties of soil sity index (Figure S4) implied that the Simpson and Shannon
underwent a significant change (Table 1). These findings once indexes of bacteria and fungi were significantly positively cor-
again illustrated the importance of balanced fertilizer applica- related with SOM and available N and that they had a negative
tion to China’s food security and soil health. Compared to the correlation with pH and Olsen-P. Previous studies showed that
PK treatment, the pH values in the NK, NP, and NPK treat- the adverse impact of soil acidification in paddy soil caused
ments had a varying amount of reduction, especially the NPK by the superfluous application of N or NPK fertilizer on soil
treatment. Previous research had shown that the application microbial diversity far exceeded the positive effect of fertiliza-
of large amounts of N fertilizer could cause soil acidification. tion (Ullah et al., 2020). Therefore, applying organic fertilizer
There are two factors involved the soil acidification: (1) the in combination with NPK chemical fertilizer was reported
exchange of NH4 + -N with H+ on root surface and (2) the nitri- more effective than providing NPK fertilizer alone in future
fication of unutilized NH4 + -N in soil, which will produce H+ (Liu, Chen, et al., 2020; Wang et al., 2017).
and lower the soil pH and microbial population (Yan et al.,
2020). There, the combination of organic and inorganic fertil-
izers can reduce soil acidification, and thus improve microbial 4.4 Effect of long-term fertilization on the
diversity. composition of soil microbial community

The predominant bacterial phyla (Figure 2) were Acidobacte-


4.3 Effects of long-term fertilization on soil ria, Proteobacteria, and Actinobacteria, and this is consistent
microbial alpha diversity with the phyla reported by Wang et al. (2018) and Guo et al.
(2019). Proteobacteria, Nitrospirae, Firmicute, and Acti-
Alpha diversity of soil bacterial communities plays an impor- nobacteria (i.e., R-strategist) are considered as copiotrophic
tant role in the sustainable development of agriculture (Eo & groups, while Acidobacteria, Actinobacteria, Chloroflexi, and
Park, 2016; Guo et al., 2020). The NK treatment had the high- Planctomycetes were typical oligotrophic (i.e., K-strategist)
est richness and diversity index, and the PK treatment had the bacteria (Fierer et al., 2012). Sun et al. (2015) found that
lowest values, which was consistent with the findings of Ling the relative abundance of Proteobacteria was significantly
et al. (2017). However, the results of Islam et al. (2012) indi- positively correlated with soil carbon and N content, which
cated that the addition of NPK fertilizer had little effect on is consistent with our results (Figure 6). Regarding the
microbial activity with long-term fertilization. This contradic- discrepancy, the relative abundances of Bacteroidetes, Gem-
tion may be related to the level of soil fertility and the duration matimonadetes, and Firmicutes without P fertilization (NK
of the field experiment. From the fungal alpha diversity, the treatment) increased markedly, while Bacteroidetes, Gemma-
change tendency in different treatments was like that of bac- timonadetes, and Firmicutes without N and K fertilization
teria. Fungi occupies a large proportion of the soil microbial decreased significantly. These results showed that different
community, and there are many fungi distributed in the soil nutrient fertilizer led to major changes in the relative abun-
environment. They can decompose organic matter and can dance of particular bacterial communities (Miransari, 2011).
be used as a significant biological indicator to evaluate the Both long-term and short-term studies have found that the
soil quality. For crops, they have an effective nutrient supply abundance of Acidobacteria decreased significantly with the
function; for example, the arbuscular mycorrhizal fungi could increase of N fertilizer application, which was closely asso-
directly activate mineral-N and K and transfer to host plant ciated with soil pH value (Liu, Li, et al., 2020), while, in
(Miransari, 2011). As shown in Table 1, the available P limi- our study, the relative abundance of Acidobacteria, serving
tation in our experiment was remarkable, which caused great as typically oligotrophic bacteria in the recommended NPK
changes in the composition of P-related microbial species. fertilization, was higher than that in other fertilization, which
Bulgarelli et al. (2019) found that the alpha diversity index of was in contradiction with previous research (Eo & Park, 2016;
the fungal community was significantly decreased after con- Islam et al., 2012). This may be related to the increase in the
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 219

special subgroup function of Acidobacteria, which, in terms the soil microbial diversity, whereas the PK and NP treatments
of genus, were unclassified Subgroup 3, unclassified Sub- had a contrary impact, compared with the NPK treatment. The
group 6, unclassified Subgroup 17, and uncultured Desul- unbalanced fertilization treatment increased the abundance of
fovirga sp. (Figure 5). Additionally, balanced fertilization oligotrophic bacteria, such as Actinobacteria, Chloroflexi, and
could stimulate the propagation of Acidobacteria phylum Planctomycetes, and the relative abundance of Basidiomy-
community, thereby activating the insoluble mineral ions in cota and Mortierellomycota. The application of NPK fertilizer
the soil (Sun et al., 2015; Wang et al., 2017). increased the abundance of the copiotrophic bacteria, such
In the fungi groups, Basidiomycota (oligotrophic) and as the Firmicutes, Gemmatimonadetes, and Proteobacteria
Ascomycota (eutrophic) were the dominant phyla communi- phyla, and the relative abundance of Ascomycota. Available
ties. Figure 2 shows that the NK fertilization significantly K, pH, and slowly available K were closely related to alter-
reduced the abundance of Ascomycota and increased the rel- ations in the composition of the soil bacterial community,
ative abundance of Basidiomycota and Mortierellomycota, while, Olsen-P, pH, and slowly available K were significantly
which is in accordance with the findings of Wang et al. (2018). correlated with the fungal community. Soil properties altered
The relative abundances of Basidiomycota, Mortierellomy- by fertilization management had a more direct effect on crop
cota, and Rozellomycota in the NP fertilization decreased yield than the community composition and the alpha diversity
obviously compared with the NPK fertilization. In compari- of bacterial and fungal groups.
son, Olpadiomycota and Glomeromycota in the PK treatment
increased by varying degrees (Figure 2). This again proves AU T H O R C O N T R I B U T I O N S
that the unbalanced fertilizer input leads to an increase in olig- Jifu Li: Project administration; writing – original draft;
otrophic fungi and a decrease in eutrophic fungi. Particularly, writing – review & editing. Guoyu Gan: Data curation;
according to the results of the cluster analysis at the genus investigation. Qianqian Zeng: Investigation.
level, there were significant differences in the distribution of
fungal species among the four treatments (Figure 3), which
AC K N OW L E D G M E N T S
may be closely related to soil nutrient status and response to
This work was supported in part by the Open-end Fund
crop yield. Fan et al. (2020) found that the microbial resistance
of Engineering Research Center of Ecology and Agricul-
indirectly drives the effects of nutrient fertilization on plant
tural Use of Wetland, Ministry of Education (KF2020015)
production. This phenomenon was also found in the results of
and the Open Fund of Key Laboratory of Fertilizer Uti-
RDA, that is, Olsen-P and slowly available K had the positive
lization, Ministry of Agriculture and Rural Affairs, China
influence on the fungal community and rice or rapeseed yield
(KLFAW201901). The authors earnestly thank the editors and
(Figures 6 and 7). The reasons the microbe influences the
crop production may be that (1) balanced fertilization could anonymous reviewers for their work.
facilitate plants acquiring more nutrients and less competi-
tion from microbial species in soil and (2) abundant nutrient CONFLICT OF INTEREST
could higher benefit bacteria and lower relative abundance of The authors declare no conflict of interest.
potential fungal plant pathogens (Fan et al., 2020; Li & Zhong,
2021). Our study also indicated that soil properties altered ORCID
by fertilization management had a more direct effect on crop Jifu Li https://orcid.org/0000-0002-3639-9434
yield than the community composition and the alpha diversity
REFERENCES
of bacterial and fungal groups. Therefore, the balanced fertil-
ization with currently recommended NPK fertilization could Bai, Y., Wang, L., Lu, Y., Yang, L., Zhou, L., Ni, L., & Cheng, M. (2015).
Effects of long-term full straw return on yield and potassium response
enhance the paddy-upland rotation production as well as the
in wheat-maize rotation. Journal of Integrative Agriculture, 14, 2467–
soil microbial population and diversity. 2476. https://doi.org/10.1016/S2095-3119(15)61216-3
Barberan, A., Ramirez, K. S., Leff, J. W., Bradford, M. A., Wall, D. H.,
& Fierer, N. (2014). Why are some microbes more ubiquitous than
5 CONCLUSIONS others? Predicting the habitat breadth of soil bacteria. Ecology Letters,
17, 794–802. https://doi.org/10.1111/ele.12282
Long-term deprivation of fertilizer input significantly influ- Bulgarelli, R. G., Silva, F. M. D., Bichara, S., Andrade, S. A. L., &
Mazzafera, P. (2019). Eucalypts and low phosphorus availability:
enced paddy-upland rotation yield and soil properties, espe-
Between responsiveness and efficiency. Plant and Soil, 445(1-2),
cially pH and available K content. Compared with NPK
349–368. https://doi.org/10.1007/s11104-019-04316-2
fertilization, non-N, -P, and -K fertilization led a decline in Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.
the average annual yield of rice and rapeseed. Rapeseed was D., Costello, E. K., Fierer, N., Pena, A. G., Goodrich, J. K., Gordon,
more sensitive to insufficient fertilizer input than rice in the J. I., Huttley, G. A., Kelley, S. T., Knights, D., Koenig, J. E., Ley, R.
same site in the rotation system. The NK treatment increased E., Lozupone, C. A., McDonald, D., Muegge, B. D., Pirrung, M., . . .
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
220 LI ET AL.

Knight, R. (2010). QIIME allows analysis of high-throughput com- analysis. Pedosphere, 27, 364–370. https://doi.org/10.1016/S1002-
munity sequencing data. Nature Methods, 7, 335–336. https://doi.org/ 0160(17)60324-2
10.1038/nmeth.f.303 Li, J. F., & Zhong, F. F. (2021). Nitrogen release and re-adsorption
Chen, M., Xu, J., Li, Z., Zhao, B., & Zhang, J. (2020). Soil physicochem- dynamics on crop straw residue during straw decomposition in an
ical properties and bacterial community composition jointly affect Alfisol. Journal of Integrative Agriculture, 20, 248–259. https://doi.
crop yield. Agronomy Journal, 112, 4358–4372. https://doi.org/10. org/10.1016/S2095-3119(20)63238-5
1002/agj2.20358 Ling, N., Chen, D. M., Guo, H., Wei, J. X., Bai, Y. F., Shen, Q. R., &
Cookson, W. R., Abaye, D. A., Marschner, P., Murphy, D. V., Stockdale, Hu, S. J. (2017). Differential responses of soil bacterial communities
E. A., & Goulding, K. W. T. (2005). The contribution of soil organic to long-term N and P inputs in a semi-arid steppe. Geoderma, 292,
matter fractions to carbon and nitrogen mineralization and microbial 25–33. https://doi.org/10.1016/j.geoderma.2017.01.013
community size and structure. Soil Biology & Biochemistry, 37(9), Liu, C., Li, L. L., Xie, J. H., Coulter, J. A., Zhang, R. Z., Luo, Z. Z.,
1726–1737. https://doi.org/10.1016/j.soilbio.2005.02.007 Cai, L. Q., Wang, L. L., & Gopalakrishnan, S. (2020). Soil bac-
Dai, Z., Liu, G., Chen, H., Chen, C., Wang, J., Ai, S., Wei, D., Li, D., terial diversity and potential functions are regulated by long-term
Ma, B., Tang, C., Brookes, P. C., & Xu, J. (2020). Long-term nutrient conservation tillage and straw mulching. Microorganisms, 8(6), 836.
inputs shift soil microbial functional profiles of phosphorus cycling https://doi.org/10.3390/microorganisms8060836
in diverse agroecosystems. The ISME Journal, 14, 757–770. https:// Liu, J., Chen, X. F., Liu, M., Wu, M., Wang, B. R., Cai, Z. H., Zhang,
doi.org/10.1038/s41396-019-0567-9 T. L., & Li, Z. P. (2020). Effects of long-term fertilization on bacte-
Ding, J. L., Jiang, X., Ma, M. C., Zhou, B. K., Guan, D. W., Zhao, B. rial community in upland red soil [Chinese with English abstract].
S., Zhou, J., Cao, F. M., Li, L., & Li, J. (2016). Effect of 35 years Acta Pedologica Sinica, 57(2), 468–478. https://doi.org/10.11766/
inorganic fertilizer and manure amendment on structure of bacterial trxb201810090507
and archaeal communities in black soil of northeast China. Applied Lu, S., Lepo, J. E., Song, H. X., Guan, C. Y., & Zhang, Z. H. (2018).
Soil Ecology, 105, 187–195. https://doi.org/10.1016/j.apsoil.2016.04. Increased rice yield in long-term crop rotation regimes through
010 improved soil structure, rhizosphere microbial communities, and
Edgar, R. C. (2010). Search and clustering orders of magnitude faster nutrient bioavailability in paddy soil. Biology and Fertility of Soils,
than BLAST. Bioinformatics, 26(19), 2460–2461. https://doi.org/10. 54(4), 909–923. https://doi.org/10.1007/s00374-018-1315-4
1093/bioinformatics/btq461 Miransari, M. (2011). Soil microbes and plant fertilization. Applied
Eo, J., & Park, K. C. (2016). Long-term effects of imbalanced fertiliza- Microbiology and Biotechnology, 92, 875–885. https://doi.org/10.
tion on the composition and diversity of soil bacterial community. 1007/s00253-011-3521-y
Agriculture Ecosystems & Environment, 231, 176–182. https://doi. Mooshammer, M., Wanek, W., Hammerle, I., Fuchslueger, L., Hofhansl,
org/10.1016/j.agee.2016.06.039 F., Knoltsch, A., Schnecker, J., Takriti, M., Watzka, M., Wild, B.,
Fan, K. K., Delgado-Baquerizo, M., Guo, X. S., Wang, D. Z., Zhu, Y. G., Keiblinger, K. M., Zechmeister-Boltenstern, S., & Richter, A. (2014).
& Chu, H. Y. (2020). Microbial resistance promotes plant pro-duction Adjustment of microbial nitrogen use efficiency to carbon: Nitrogen
in a four-decade nutrient fertilization experiment. Soil Biology imbalances regulate soil nitrogen cycling. Nature Communications, 5,
& Biochemistry, 141, 107679. http://doi.org/10.1016/j.soilbio.2019. 3694. https://doi.org/10.1038/ncomms4694
107679 Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., O’Hara, R. B.,
Fan, K. K., Weisenhorn, P., Gilbert, J. A., & Chu, H. Y. (2018). Wheat Simpson, G. L., Solymos, P., Stevens, H., & Wagner, H. H. (2010).
rhizosphere harbors a less complex and more stable microbial co- Multivariate analysis of ecological communities in R: package
occurrence pattern than bulk soil. Soil Biology & Biochemistry, 125, “vegan”. Vegan Community Ecology Package.
251–260. https://doi.org/10.1016/j.soilbio.2018.07.022 Radhakrishnan, S., & Varadharajan, M. (2016). Status of microbial
Fierer, N., Lauber, C. L., Ramirez, K. S., Zaneveld, J., Bradford, M. A., diversity in agroforestry systems in Tamilnadu, India. Journal of
& Knight, R. (2012). Comparative metagenomic, phylogenetic and Basic Microbiology, 56(6), 662–669. https://doi.org/10.1002/jobm.
physiological analyses of soil microbial communities across nitrogen 201500639
gradients. ISME Journal, 6(5), 1007–1017. https://doi.org/10.1038/ Reed, S. C. (2017). Disentangling the complexities of how legumes
ismej.2011.159 and their symbionts regulate plant nitrogen access and storage. New
Guo, Z. B., Liu, H., Wan, S. X., Hua, K. K., Wang, D. Z., Guo, X. S., & Phytologist, 213(2), 478–480. https://doi.org/10.1111/nph.14390
He, C. L. (2019). Fertilization practice changes rhizosphere micro- Schloss, P. D., Westcott, S. L., Ryabin, T., Hall, J. R., Hartmann, M.,
bial community structure in the agroecosystem. Annals of Applied Hollister, E. B., Lesniewski, R. A., Oakley, B. B., Parks, D. H.,
Biology, 174(2), 123–132. https://doi.org/10.1111/aab.12478 Robinson, C. J., Sahl, J. W., Stres, B., Thallinger, G. G., Van Horn,
Guo, Z. B., Wan, S. X., Hua, K. K., Yin, Y., Chu, H. Y., Wang, D. Z., D. J., & Weber, C. F. (2009). Introducing mothur: Open-source,
& Guo, X. S. (2020). Fertilization regime has a greater effect on soil platform-independent, community-supported software for describing
microbial community structure than crop rotation and growth stage and comparing microbial communities. Applied and Environmental
in an agroecosystem. Applied Soil Ecology, 149, 103510. https://doi. Microbiology, 75, 7537–7541. https://doi.org/10.1128/aem.01541-09
org/10.1016/j.apsoil.2020.103510 Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett,
Islam, M. R., Sultana, T., Cho, J.-C., Joe, M. M., & Sa, T. M. W. S., & Huttenhower, C. (2011). Metagenomic biomarker discovery
(2012). Diversity of free-living nitrogen-fixing bacteria associated and explanation. Genome Biology, 12, R60. https://doi.org/10.1186/
with Korean paddy fields. Annals of Microbiology, 62(4), 1643–1650. gb-2011-12-6-r60
https://doi.org/10.1007/s13213-012-0421-z Sun, R. B., Zhang, X. X., Guo, X. S., Wang, D. Z., & Chu, H. Y. (2015).
Li, H., Lu, J., Ren, T., Li, X., & Cong, R. (2017). Nutrient efficiency Bacterial diversity in soils subjected to long-term chemical fertiliza-
of winter oilseed rape in an intensive cropping system: A regional tion can be more stably maintained with the addition of livestock
14350645, 2023, 1, Downloaded from https://acsess.onlinelibrary.wiley.com/doi/10.1002/agj2.21264 by Cochrane Costa Rica, Wiley Online Library on [19/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LI ET AL. 221

manure than wheat straw. Soil Biology & Biochemistry, 88(4), 9–18. Yan, G., Xing, Y., Han, S., Zhang, J., Wang, Q., & Mu, C. (2020). Long-
https://doi.org/10.1016/j.soilbio.2015.05.007 time precipitation reduction and nitrogen deposition increase alter
Tang, J., Huang, F., Li, J. F., Xiao, K., Zou, J. L., Zhu, J. Q., & Qin, Y. P. soil nitrogen dynamic by influencing soil bacterial communities and
(2018). Different adaptations of winter rapes and weeds responding functional groups. Pedosphere, 30, 363–377. https://doi.org/10.1016/
to long-term balanced fertilization [Chinese with English abstract]. S1002-0160(19)60834-9
Soils, 50(2), 291–297. https://doi.org/10.13758/j.cnki.tr.2018.02.010 Yousaf, M., Li, J., Lu, J., Ren, T., Cong, R., Fahad, S., &
Tian, H., Wang, H., Hui, X. L., Wang, Z. H., Drijber, R. A., & Liu, J. Li, X. (2017). Effects of fertilization on crop production and
S. (2017). Changes in soil microbial communities after 10 years of nutrient-supplying capacity under rice-oilseed rape rotation sys-
winter wheat cultivation versus fallow in an organic-poor soil in the tem. Scientific Reports, 7, 1270. https://doi.org/10.1038/s41598-017-
loess plateau of China. PLoS ONE, 12(9), e0184223. https://doi.org/ 01412-0
10.1371/journal.pone.0184223 Yu, Z. H., Hu, X. J., Wei, D., Liu, J. J., Zhou, B. K., Jin, J., Liu, X.
Tian, Z., Ji, Y. H., Sun, L. X., Xu, X. L., Fan, D. L., Zhong, H., Liang, B., & Wang, G. H. (2019). Long-term inorganic fertilizer use influ-
Z., & Ficsher, G. (2018). Changes in production potentials of rapeseed ences bacterial communities in Mollisols of Northeast China based on
in the Yangtze River Basin of China under climate change: A multi- high-throughput sequencing and network analyses. Archives of Agron-
model ensemble approach. Journal of Geographical Sciences, 28(11), omy and Soil Science, 65(10), 1331–1340. https://doi.org/10.1080/
1700–1714. https://doi.org/10.1007/s11442-018-1538-1 03650340.2018.1563685
Ullah, S., He, P., Ai, C., Zhao, S. C., Ding, W. C., Song, D. L., Zhang, J. Zeng, J., Liu, X. J., Song, L., Lin, X. G., Zhang, H. Y., Shen, C. C., &
J., Huang, S. H., Abbas, T., & Zhou, W. (2020). How do soil bacterial Chu, H. Y. (2016). Nitrogen fertilization directly affects soil bacterial
diversity and community composition respond under recommended diversity and indirectly affects bacterial community composition. Soil
and conventional nitrogen fertilization regimes? Microorganisms, 8, Biology & Biochemistry, 92, 41–49. https://doi.org/10.1016/j.soilbio.
1193. https://doi.org/10.3390/microorganisms8081193 2015.09.018
Wang, J. C., Song, Y., Ma, T. F., Raza, W., Li, J., Howland, J. G., Zhang, Z., Cong, R. H., Ren, T., Li, H., Zhu, Y., & Lu, J. W. (2020).
Huang, Q. W., & Shen, Q. R. (2017). Impacts of inorganic and Optimizing agronomic practices for closing rapeseed yield gaps under
organic fertilization treatments on bacterial and fungal communities intensive cropping systems in China. Journal of Integrative Agri-
in a paddy soil. Applied Soil Ecology, 112, 42–50. https://doi.org/10. culture, 19(5), 1241–1249. https://doi.org/10.1016/S2095-3119(19)
1016/j.apsoil.2017.01.005 62748-6
Wang, Y., Zhao, X., Guo, Z. Y., Jia, Z. J., Wang, S. Q., & Ding, K. (2018).
Response of soil microbes to a reduction in phosphorus fertilizer in
rice-wheat rotation paddy soils with varying soil P levels. Soil and
Tillage Research, 181, 127–135. https://doi.org/10.1016/j.still.2018. S U P P O RT I N G I N F O R M AT I O N
04.005 Additional supporting information can be found online in the
Wasaki, J., Sakaguchi, J., Yamamura, T., Ito, S., Shinano, T., Osaki, Supporting Information section at the end of this article.
M., & Kandeler, E. (2018). P and N deficiency change the relative
abundance and function of rhizosphere microorganisms during cluster
root development of white lupin (Lupinus albus L.). Soil Science and
Plant Nutrition, 64(6), 686–696. https://doi.org/10.1080/00380768.
How to cite this article: Li, J., Gan, G., & Zeng, Q.
2018.1536847
(2023). Impacts of long-term fertilization on crop
Yan, F., Sun, Y., Hui, X., Jiang, M., Xiang, K., Wu, Y., Zhang, Q., Tang,
Y., Yang, Z., Sun, Y., & Jun, M. (2019). The effect of straw mulch yield and microbial communities under rice–rapeseed
on nitrogen, phosphorus and potassium uptake and use in hybrid rotation. Agronomy Journal, 115, 208–221.
rice. Paddy Water Environment, 17, 23–33. https://doi.org/10.1007/ https://doi.org/10.1002/agj2.21264
s10333-018-0680-9

You might also like